Based on Quantum Topological Stabilizer Color Code Morphism Neural Network Decoder

Solving for quantum error correction remains one of the key challenges computing. Traditional decoding methods are limited by computing power and data scale, which restrict efficiency color codes. There many that have been suggested to solve this problem. Machine learning is considered most suitable solutions task code. We project code onto surface code, use deep Q network iteratively train process obtain relationship between inversion rate logical trained model performance correction. Our results show through unsupervised learning, when iterative training at least 300 times, a self-trained can improve accuracy 96.5%, speed about 13.8% higher than traditional algorithm. numerically our method achieve fast prediction after better threshold.